If the large language models didn t refuse to answer, they would oftentimes hallucinate and come up with figures and facts that are not in the SEC filings.
Patronus AI releases SimpleSafetyTests, a new diagnostic test suite for large language models that found critical safety gaps and vulnerabilities across popular AI systems like ChatGPT.
Patronus AI has uncovered limitations in AI models when analyzing SEC filings. The findings of the study shed light on the challenges faced by AI models.